Q: op-amp to protect DSP board?

Hello,

I recently got an ADUC831 microconverter board, and am interested in
experimenting with DSP applied to an audio input and/or output. A
friend who is knowledgeable about these boards recommended "passing
audio input through an op-amp to protect the board."

His meaning appears to be that I should connect my 1/8 in. jack to an
op-amp, and the op-amp to the ADC input, in such a way that if I
should inadvertantly overdrive the board, the ($1.50) op-amp will blow
first and spare the ($70) board. IOW, that I should use the op-amp as
a fancy fuse, and possibly as a clipper circuit at the same time.

Is that a standard practice? Can anyone offer pointers on using an
op-amp in this way? Alternately, can anyone recommend a better
approach?

Thanks!
Len.

 

[Jerry Avins]

lbudney@pobox.com wrote:

Hello,

I recently got an ADUC831 microconverter board, and am interested in
experimenting with DSP applied to an audio input and/or output. A
friend who is knowledgeable about these boards recommended "passing
audio input through an op-amp to protect the board."

His meaning appears to be that I should connect my 1/8 in. jack to an
op-amp, and the op-amp to the ADC input, in such a way that if I
should inadvertantly overdrive the board, the ($1.50) op-amp will blow
first and spare the ($70) board. IOW, that I should use the op-amp as
a fancy fuse, and possibly as a clipper circuit at the same time.

Is that a standard practice? Can anyone offer pointers on using an
op-amp in this way? Alternately, can anyone recommend a better
approach?

Thanks!
Len.

First off, your approach would mean modifying the board in a way that I
would avoid. If I felt that I needed that protection, I would connect my
signal to an off-board op-amp and connect its output to the jack. If
your source is a microphone, you don't need any protection at all. For
other sources, a series resistor might be perfectly adequate. You might
also consider back-to-back Zener diodes.

Jerry
--
Engineering is the art of making what you want from things you can get.
ŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻ

 

[Len Budney]

Jerry,

Thanks for your reply!

Jerry Avins <jya@ieee.org> wrote:

lbudney@pobox.com wrote:

[using an op-amp to protect a DSP board]

First off, your approach would mean modifying the board in a way that I
would avoid.

If I implied that, I apologize. The suggestion was to construct a
board with an op-amp, and then use wire-wrap to connect that to the
board.

If I felt that I needed that protection, I would connect my
signal to an off-board op-amp and connect its output to the jack.

A big part of my question is, "When do you feel you need it?" How
about when connecting standard audio devices, such as walkmen or
radios, to the board's input? Or speakers to its output?

If your source is a microphone, you don't need any protection at all. For
other sources, a series resistor might be perfectly adequate. You might
also consider back-to-back Zener diodes.

What circuit are you envisioning? I thought of using a zener between
the input pin and the ground pin as a voltage regulator, but I'm not
sure what "back-to-back" means in this context.

Upon further searching, I believe my friend's suggestion was to use
the op-amp as a voltage follower. Does that by itself do any good?
Would a voltage follower preceded by a zener (bridging the + input to
the board's ground) be better? Or in that case, is the zener the only
useful component anyway?

I'm close to designing a circuit as follows, but have no idea if it's
on the right track:

1) One line from the jack to the ground pin (using wire-wrap).
2) The other line to two resistors and a zener, forming a T.
3) The zener bridges the jack's leads.
4) Output of the T network to the + input of the op-amp.
5) Op-amp output to the - input and to the board's input.

Between the output and the - input, I considered tweaking with a
resistor, if necessary, so that the audio device's max output will
arrive at maybe 90% of the board's max allowed input. Critique?

Thanks!
Len.

 

[Al Clark]

lbudney@pobox.com (Len Budney) wrote in
news:9213c87c.0312140524.68f0af4b@posting.google.com:

Chances are very good that you don't need anything. Without seeing the
front end circuit, this is a guess:

If it makes you feel better, add a 1000 Ohm series resistor to the input.
I am assuming that the input impedance might be about 10K. This will
attenuate the signal by about 1dB.

--
Al Clark
Danville Signal Processing, Inc.
--------------------------------------------------------------------
Purveyors of Fine DSP Hardware and other Cool Stuff
Available at http://www.danvillesignal.com

Jerry,

Thanks for your reply!

Jerry Avins <jya@ieee.org> wrote:
lbudney@pobox.com wrote:

[using an op-amp to protect a DSP board]

First off, your approach would mean modifying the board in a way that
I would avoid.

If I implied that, I apologize. The suggestion was to construct a
board with an op-amp, and then use wire-wrap to connect that to the
board.

If I felt that I needed that protection, I would connect my
signal to an off-board op-amp and connect its output to the jack.

A big part of my question is, "When do you feel you need it?" How
about when connecting standard audio devices, such as walkmen or
radios, to the board's input? Or speakers to its output?

If your source is a microphone, you don't need any protection at
all. For other sources, a series resistor might be perfectly
adequate. You might also consider back-to-back Zener diodes.

What circuit are you envisioning? I thought of using a zener between
the input pin and the ground pin as a voltage regulator, but I'm not
sure what "back-to-back" means in this context.

Upon further searching, I believe my friend's suggestion was to use
the op-amp as a voltage follower. Does that by itself do any good?
Would a voltage follower preceded by a zener (bridging the + input to
the board's ground) be better? Or in that case, is the zener the only
useful component anyway?

I'm close to designing a circuit as follows, but have no idea if it's
on the right track:

1) One line from the jack to the ground pin (using wire-wrap).
2) The other line to two resistors and a zener, forming a T.
3) The zener bridges the jack's leads.
4) Output of the T network to the + input of the op-amp.
5) Op-amp output to the - input and to the board's input.

Between the output and the - input, I considered tweaking with a
resistor, if necessary, so that the audio device's max output will
arrive at maybe 90% of the board's max allowed input. Critique?

Thanks!
Len.

--
Al Clark
Danville Signal Processing, Inc.
--------------------------------------------------------------------
Purveyors of Fine DSP Hardware and other Cool Stuff
Available at http://www.danvillesignal.com

 

[Jerry Avins]

Len Budney wrote:

Jerry,

Thanks for your reply!

Jerry Avins <jya@ieee.org> wrote:

lbudney@pobox.com wrote:

[using an op-amp to protect a DSP board]

First off, your approach would mean modifying the board in a way that I
would avoid.


If I implied that, I apologize. The suggestion was to construct a
board with an op-amp, and then use wire-wrap to connect that to the
board.

No need even to wire wrap. Use a piece of shielded cable with an
eighth-inch plug on the end. Anyhow, shielded cable is hard to wire
wrap. (With driver impedance low enough, twisted pair might suffice.)

If I felt that I needed that protection, I would connect my
signal to an off-board op-amp and connect its output to the jack.


A big part of my question is, "When do you feel you need it?" How
about when connecting standard audio devices, such as walkmen or
radios, to the board's input? Or speakers to its output?

Use the back of an envelope. The jack is certainly able to handle a
typical "line out" signal. That's at least 1 V RMS. It would be poor
design if 3-volt peaks could damage it. Look at the codec spec sheet. My
guess is that driving the input even to the rails won't hurt if the
current is kept low enough. Internally, it's probably clamped to the
rails, but maybe not. Look and see. What is the operating impedance of
the input to the chip? My guess is that it's in the order of a megohm.

Speakers should be connected through a voltage divider. At very high
outputs, the audio may be distorted, but a modest output signal
overrides the output-stage noise present in all but the best amplifiers.
Making the impedance of the voltage might be all the additional
protection needed. Too high an impedance, whether of a divider or a
stopper resistor, will degrade the highs.

A CD player rarely operates on more than 6 volts, so even with an
H-bridge driver, the headphone output won't exceed 5 V peak. (10 V P-P:
ouch!) The line out will be lower voltage and less distortion. Use that.
Assuming that the input is clamped to 5-volts and ground and the source
voltage can go as high as ten volts peak, 1 10 K series resistor will
limit the input current to 1 ma peak (but you need a bigger resistor to
withstand shorts to the power line).

If your source is a microphone, you don't need any protection at all. For
other sources, a series resistor might be perfectly adequate. You might
also consider back-to-back Zener diodes.

What circuit are you envisioning? I thought of using a zener between
the input pin and the ground pin as a voltage regulator, but I'm not
sure what "back-to-back" means in this context.

A Zener breaks down in the reverse direction. In the forward direction,
its characteristics are similar to any other semiconductor diode. To get
a double knee, you need to connect two anode to anode or cathode to
cathode.

Upon further searching, I believe my friend's suggestion was to use
the op-amp as a voltage follower. Does that by itself do any good?

Not for large inputs. It may isolate the board from static discharges.
But it might not.

Would a voltage follower preceded by a zener (bridging the + input to
the board's ground) be better? Or in that case, is the zener the only
useful component anyway?

Probably.

I'm close to designing a circuit as follows, but have no idea if it's
on the right track:

1) One line from the jack to the ground pin (using wire-wrap).
2) The other line to two resistors and a zener, forming a T.
3) The zener bridges the jack's leads.
4) Output of the T network to the + input of the op-amp.
5) Op-amp output to the - input and to the board's input.

Between the output and the - input, I considered tweaking with a
resistor, if necessary, so that the audio device's max output will
arrive at maybe 90% of the board's max allowed input. Critique?

A decent board should be capable of withstanding moderate overload. 90%
of "maximum allowable" is probably OK. That ought to be nearly double
the maximum for linear operation. The circuit you describe seems liable
to hum pickup. Allow for shielding to be added later if it's needed.
That means that flying wires won't do. You want the noise floor to be
set by quantization. If you aren't careful, hum and noise from the
preamp will dominate.

Thanks!

You're welcome. Good luck.

Jerry
--
Engineering is the art of making what you want from things you can get.
ŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻ

 

[Ban]

Len Budney wrote:
||
|| A big part of my question is, "When do you feel you need it?" How
|| about when connecting standard audio devices, such as walkmen or
|| radios, to the board's input? Or speakers to its output?
||
||| If your source is a microphone, you don't need any protection at
||| all. For other sources, a series resistor might be perfectly
||| adequate. You might also consider back-to-back Zener diodes.
||
|| What circuit are you envisioning? I thought of using a zener between
|| the input pin and the ground pin as a voltage regulator, but I'm not
|| sure what "back-to-back" means in this context.
||
|| Upon further searching, I believe my friend's suggestion was to use
|| the op-amp as a voltage follower. Does that by itself do any good?
|| Would a voltage follower preceded by a zener (bridging the + input to
|| the board's ground) be better? Or in that case, is the zener the only
|| useful component anyway?
||
|| I'm close to designing a circuit as follows, but have no idea if it's
|| on the right track:
||
|| 1) One line from the jack to the ground pin (using wire-wrap).
|| 2) The other line to two resistors and a zener, forming a T.
|| 3) The zener bridges the jack's leads.
|| 4) Output of the T network to the + input of the op-amp.
|| 5) Op-amp output to the - input and to the board's input.
||
|| Between the output and the - input, I considered tweaking with a
|| resistor, if necessary, so that the audio device's max output will
|| arrive at maybe 90% of the board's max allowed input. Critique?
||
|| Thanks!
|| Len.

That thing with the opamp is not a protection at all, who wants to blow the
opamp, it needed to be replaced and probably it could still blow the board.
You really need a good protection if you have some automobile environment.
There short transients up to 150V can occurr. The simple circuit down will
do that. The fuse can be dropped if you use a 10k or don't have a continuous
high voltage such as mains to protect from. There should be some big
electrolytic caps on the board (100uF) or a 5.6V zener to protect the supply
in case it is 5V.


VCC VCC
+ --
| \
- BAS70-04 /\ 5.6V Zener
0.03A fast ^ dual schottky -- once only
____ ___ | |
o--|_--_|---|___|--+-----o ===
fuse 1k...10k | port pin GND
-
^
|
each port ===
GND

If you want an even better protection without much current flowing you could
also use a diode ring which will continuously protect even to mains without
a series resistor, but it might induce a few millivolts of offset.
Eventually only the resistors will burn out. Again both +/-5V need a zener
across like above.

+5V
|
.-.
| |10k
| |
'-'
|
+-|<-+->|-+
in | | port
o-------+ 1N4936 +---+--o
| | |
+->|-+-|<-+ -
| ^4148
.-. |
| |10k ===
| | GND
'-'
|
-5V
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de


--
ciao Ban
Bordighera, Italy
electronic hardware designer